Economic selection index development for Beefmaster cattle I: Terminal breeding objective

University of Nebraska - Lincoln

DigitalCommons@University of Nebraska - Lincoln

Faculty Papers and Publications in Animal Science

Animal Science Department

2017

Economic selection index development for

Beefmaster cattle I: Terminal breeding objective

Kathleen P. Ochsner

University of Nebraska-Lincoln, [email protected]

M. D. MacNeil

Delta G, Miles City, Montana

Matthew L. Spangler

University of Nebraska-Lincoln, [email protected]

Follow this and additional works at:

http://digitalcommons.unl.edu/animalscifacpub

Part of the

Genetics and Genomics Commons

, and the

Meat Science Commons

This Article is brought to you for free and open access by the Animal Science Department at DigitalCommons@University of Nebraska - Lincoln. It has been accepted for inclusion in Faculty Papers and Publications in Animal Science by an authorized administrator of DigitalCommons@University of Nebraska - Lincoln.

Ochsner, Kathleen P.; MacNeil, M. D.; and Spangler, Matthew L., "Economic selection index development for Beefmaster cattle I: Terminal breeding objective" (2017).Faculty Papers and Publications in Animal Science. 988.

1063

INTRODUCTION

The main source of long-term profitability for a beef cattle operation lies in its production efficien -cy, which can be improved through genetic selection.

Traditionally, EBV have been the genetic tools used to select breeding livestock. While EBV are a sound se -lection tool, a drawback is that they represent genetic merit in only one trait while in reality multiple traits

influence an animal’s value (Hazel, 1943). With EBV

as a sole selection tool, producers are left to individu-ally determine their optimal use and ultimately the eco-nomic importance of each trait (Bourdon, 1998). Hazel

and Lush (1942) and Hazel (1943) first introduced the

concept of combining genetic evaluation and econom-ics through selection index theory. Since then, selection

indices have been implemented in the beef industry and are the recommended method of multi-trait selection in animal populations (Falconer and Mackay, 1996). Currently, Beefmaster Breeders United (BBU) reports

ten EBV, but provides no tool for multi-trait selection.

The objective of this study was to develop a selection index for terminal purpose Beefmaster cattle to increase

profitability of commercial enterprises and facilitate ge -netic improvement of the Beefmaster breed.

MATERIALS AND METHODS

Animal Care and Use Committee approval was not obtained for this study given that the data were simulated.

Defining the Breeding Objective

The breeding objective for development of the

terminal index was to increase profitability of an op -eration where all calves were born from mature cows, retained through the feedlot phase and sold on a grid-based pricing system. The 5 objective traits consid-ered for the terminal index included HCW, marbling

Beefmaster cattle I: Terminal breeding objective

K. P. Ochsner,* M. D. MacNeil,†‡ R. M. Lewis,* and M. L. Spangler*2 *Department of Animal Science, University of Nebraska, Lincoln 68583; †Delta G, Miles City, Montana 59301 and ‡University of the Free State, Bloemfontein, South Africa ABSTRACT: The objective of this study was to

develop an economic selection index for Beefmaster cattle in a terminal production system where bulls are mated to mature cows with all resulting progeny har-vested. National average prices from 2010 to 2014 were used to establish income and expenses for the system. Phenotypic and genetic parameter values among the selection criteria and goal traits were obtained from lit-erature. Economic values were estimated by simulating 100,000 animals and approximating the partial

deriva-tives of the profit function by perturbing traits one at a

time, by 1 unit, while holding the other traits constant at

their respective means. Relative economic values (REV)

for the terminal objective traits HCW, marbling score (MS), ribeye area (REA), 12th–rib fat (FAT), and feed intake (FI) were 91.29, 17.01, 8.38, -7.07, and -29.66,

respectively. Consequently, improving the efficiency of beef production is expected to impact profitability great -er than improving carcass m-erit alone. The accuracy of the index lies between 0.338 (phenotypic selection) and 0.503 (breeding values known without error). The application of this index would aid Beefmaster breed-ers in their sire selection decisions, facilitating genetic improvement for a terminal breeding objective.

Key words:

beef cattle, selection index, terminal objective

©

2017 American Society of Animal Science. All rights reserved

. J. Anim. Sci. 2017.95:1063–1070

doi:10.2527/jas2016.1231

1_{This work was supported by Beefmaster Breeders United and}

National Research Initiative Competitive Grant number 2011–68004– 30214 from the USDA National Institute of Food and Agriculture.

2_{Corresponding author: [email protected]}

Received November 23, 2016. Accepted December 31, 2016.

Ochsner et al. 1064

score (MS), ribeye area (REA), 12th–rib fat (FAT) and feed intake (FI), with the latter representing the only expense related phenotype among the objective traits. Choice of Selection Criteria

Ideally, the selection criteria would include all economically relevant traits in the breeding objective. However, in practice some traits in the objective are not readily observed, hence the need to use indicator traits for predicting traits that are economically rel-evant. Selection criteria should be highly correlated to the traits in the objective. Selection criteria for the

terminal index were chosen from the 10 EBV current -ly reported by BBU and were yearling weight (YW), ultrasound ribeye area (UREA), ultrasound 12th–rib fat (UFAT) and ultrasound intramuscular fat (UIMF). Estimation of Economic Values

A method to derive economic values is

par-tial differentiation of a profit equation (Hill, 1974;

Ponzoni and Newman, 1989; Forabosco et al., 2004). Identifying sources of income and expense in the beef

cattle herd enables the development of a profit equa

-tion where profit is a func-tion of income and expense

(Ponzoni and Newman, 1989). Sources of income and expense for the terminal production system were

iden-tified and the profit was simulated for 100,000 animals

using SAS 9.3 (SAS Inst. Inc., Cary, NC).

In the assumed production and marketing system, half of the calves were fed through a calf-fed system and half were fed through a yearling system. For the calf-fed system it was assumed that calves were sent

to the feedlot directly after weaning for a 211 d finish -ing period before harvest. The yearl-ing system assumed that after weaning calves entered into a 315 d growing

period prior to being sent to a feed yard for a 90 d finish -ing period. It was assumed that all replacement females were obtained from outside the herd. Income was de-rived solely from the marketing of animals for slaugh-ter on a grid based system. Phenotypes for HCW, MS, REA, and FAT were simulated from a random normal distribution with the means (SD) based on literature values of 320 (38.8) kg, 5.4 (0.9) marbling score units, 76.5 (9.3) cm2 _{and 1.2 (0.32) cm, respectively (Moser} et al., 1998; Wheeler et al., 2006). The genetic relation-ships between traits were accounted for by a Cholesky decomposition applied to the phenotypic covariance matrix between all objective traits considered.

The 5-yr (2010 to 2014) average price for steers and heifers at slaughter was obtained from the Livestock Marketing Information Center (LMIC, 2015) and used as the base price for all slaughter animals. The base

price was $3.858/kg with a SD of $0.642/kg. Premium and discount values based on yield grade (YG), qual-ity grade (QG) and HCW were obtained from United States Department of Agriculture- Agricultural Marketing Service (USDA-AMS, 2015) and are

pre-sented in Table 1. Quality grade of each carcass was

assigned based on simulated marbling score (e.g., 5.0 = Sm0 _{= Low Choice), and each animal received a} premium or discount accordingly. It was assumed that animals sent to slaughter are 30 mo of age or young-er and thus age was not considyoung-ered as a contributing

factor to QG. The standard USDA equation for yield

grade (Murphey et al., 1960) was used, after adjust-ing the intercept for the average KPH fat percentage.

The resulting equation was: YG = 3.0 + 0.984 × FAT - 0.0496 × REA + 0.00838 × HCW.Weight discounts were applied to animals for which the simulated car-cass weight was under 272 kg or over 409 kg. Carcar-cass price was calculated as the sum of base carcass price,

YG premium/discount, QG premium/discount and

weight discount (if applicable). Income for each ani-mal was calculated by multiplying the carcass price ($/ kg) by the weight of the animal in kg.

Expenses for the production system assumed in de-velopment of the terminal index were feed, veterinary labor, medicine, bedding, marketing, custom operations, fuel, repairs, processing, and yardage. A 5-yr (2010 to

2014) average and SE of prices for feedstuffs used in the

production system were calculated using information

ob-Table 1. Premiums and discounts for carcass sales based on 5-yr average (2010 to 2014)

Category Adjustment1 _($/kg)

USDA Quality Grade

Prime 0.402

Choice 0

Select -0.195 Standard -0.480

USDA Yield Grade 1.0–2.0 0.092 2.0–2.5 0.048 2.5–3.0 0.045 3.0–4.0 0 4.0–5.0 -0.228 > 5.0 -0.386 Carcass weight (kg) < 227 -0.702 227–250 -0.483 250–272 -0.061 272–409 0 409–431 -0.005 431–454 -0.006 > 454 -0.511

1_{U.S. Department of Agriculture Agricultural Marketing Service.}

tained from the USDA–National Agricultural Statistics Service (USDA-NASS, 2015). The correlation between

corn prices and other feedstuffs was included in the

simulation to ensure that the relationship between prices did not deviate from their true relationship in the indus-try. Prices for each feed ingredient were simulated from a random normal distribution as a function of the aver-age price, SE and correlation with the price of corn. Feed intake was simulated from a random normal distribution with a mean of 8.59 kg and SD of 1.09 kg (Rolfe et al., 2011). Feed costs for animals fed through the calf-fed system were simulated assuming animals were consum-ing the feedlot diet outlined in Table 2 for 211 d. Cattle in the yearling system were fed the winter yearling system diet for 198 d, the summer yearling system diet for 117 d and the feedlot diet for 90 d (Table 2).

Veterinary labor, medicine, bedding, market -ing, custom operations, fuel, repairs, process-ing, and

yardage were considered fixed while building the profit equation because they did not vary based on

the biological merit of an individual animal (Table 3).

Veterinary and medicine costs were estimated by cal -culating a 5-yr average from data provided by D. W. Gillings (Christiansen Land and Cattle Ltd., Kimball, SD, personal communication). Means and SE of other

costs including bedding, marketing, custom opera-tions, fuel, repairs, processing, and yardage were ob-tained from Barron Lopez (2013). Total cost of the production system was calculated as a sum of feed costs and other costs through all phases of production.

Calculating Selection Index Coefficients

Hazel (1943) first introduced the selection index equations to calculate index coefficients (b) for each of the selection criteria:

b = P-1_Gv

where P is a n × n matrix of the phenotypic (co)vari-ances among the n traits measured and available as selection criteria, G is a n × m matrix of the genetic (co)variances among the n selection criteria and m ob-jective traits, and v is an m × 1 vector of economic values for all objective traits. This method was used to

calculate economic index coefficients to be applied to

phenotypic measures for the terminal index. Genetic co-variances were calculated from the genetic SD and genetic correlations. Phenotypic co-variances were calculated using the phenotypic SD and phenotypic correlations between traits. The heritability, genetic variances and phenotypic variances of the objective traits and selection criteria used to calculate the P and G matrices were extracted from literature and are pre-sented in Table 4. Phenotypic correlations among the selection criteria, and genetic correlations between the selection criteria and objective traits, needed for back-calculation of the co-variances were extracted from

scientific literature and are presented in Table 5.

For an index designed for a beef breed

associa-tion index coefficients should be applied to EBV. Not

only is this more practical, but literature supports the

argument that index coefficients applied to EBV are

more accurate (Schneeberger et al., 1992; Bourdon, 1998). From a practicality standpoint, phenotypic

Table 2. Diet composition and prices of feedstuffs

based on a 5-yr average (2010 to 2014)

Ingredient Inclusion

1

(% DM) Price

2

($/kg) ($/kg)SD Correlation3

Feedlot Diet Composition

Dry-rolled corn 43.8 0.211 0.051 1.00 Wet distillers grains

+ solubles 43.8 0.200 0.048 1.00 Alfalfa hay 7.5 0.200 0.042 0.84 Urea 1.1 0.663 0.050 0.72 Limestone 1.9 0.028 0.002 0.92 Potassium 0.8 0.648 0.071 0.65 Salt 0.6 0.289 0.011 0.84 Trace minerals 0.43 0.877 0.037 0.18 Rumensin 0.03 19.575 3.915 0.40 Tylan 0.02 17.775 3.555 0.40 Vitamins 0.02 2.950 0.360 0.40 Winter Yearling System Diet Composition

Prairie hay 74 0.140 0.022 0.66 Corn 20 0.211 0.051 1.00 44% protein supplement 6 0.436 0.060 0.87 Summer Yearling System Diet Composition

Summer Grazing 75 0.105 0.022 0.90 Prairie Hay 19 0.140 0.022 0.66 Corn 5 0.211 0.051 1.00 44% protein supplement 1 0.436 0.060 0.87

1_{Based on Barron Lopez (2013).}

2_{USDA National Agricultural Statistics Service.}

3_{Correlation with the price of corn. Based on Barron Lopez (2013).}

Table 3. Price of other costs in terminal system based on average prices from 2010 to 2014

Expense object (US$/head per yr)Average cost SE of cost

Veterinary and Medicine1 _{19.220 4.464}

Bedding2 _{0.49 0.12}

Marketing2 _{10.407 3.534}

Custom Operations2 _{30.877 11.915}

Fuel2 _{53.463 10.636}

Repairs2 _{42.190 9.208} 1_{D. Gillings, Christiansen Land and Cattle Ltd., Kimball, SD, personal}

communication.

Ochsner et al. 1066

measures will rarely be available for all animals on all traits included in the selection criteria. Sex-limited traits and traits such as carcass merit cannot be mea-sured directly on all breeding animals. Initial selection decisions are often made before an animal expresses all the traits which determine its overall genetic merit. Additionally, Bourdon (1998) raised 2 serious draw-backs in applying index weighting factors to pheno-typic values for an individual. First, this method lacks accuracy because it does not incorporate information on relatives. Second, it is biased because it does not

account for genetic differences among contemporary groups. These issues can be overcome by using EBV

instead of individual phenotypic performance. Another

benefit of using index coefficients to be applied to EBV

is that the phenotypes entering into genetic evaluation

are adjusted for heterosis effects, which may be espe -cially important in a composite breed like Beefmaster. Schneeberger et al. (1992) presented a method to

calculate a vector of index coefficients to be applied to EBV for the selection criteria in the index. The equation to estimate index coefficients to be applied to EBV is:

b = G₁₁-1_G 12v

where G₁₁ is a n × n matrix of genetic (co)variances among the n selection criteria, G₁₂ is a n × m matrix of the genetic (co)variances among the n selection criteria and m objective traits and v is an m × 1 vector of

eco-nomic values for all objective traits. Index coefficients to be applied to EBV for selection criteria were calculated

using this method. For each selection index, it was

en-sured that a positive definite (co)variance matrix existed.

Estimating Index Accuracy

Following the notation of Van Vleck (1993), the

accuracies of the indices that utilize phenotypic mea-sures were calculated as:

(

)

where b'Gv represents the covariance between the in-dex and aggregate genotype, b'Pb represents the index variance, and v'Cv represents the aggregate genotype variance. C is an m × m genetic (co)variance matrix

among the objective traits.

For indices that utilize EBV as the selection crite -ria, the following equation was used to calculate the accuracy of the index:

(

)

where b'G v12 represents the covariance between the

in-dex and aggregate genotype and b'G11b represents the

index variance. The substitution of G₁₁ for P in calcu-lating the index variance is accompanied by several

as-sumptions. In presenting the index coefficient equations using EBV as the selection criteria, Schneeberger et al.

(1992) explained that G₁₁ is the genetic (co)variance matrix of the selection criteria which is assumed to be

Table 4. Genetic and phenotypic parameters for selec-tion criteria and objective traits

Trait1 _h2 σ_α2 σ_p2 _Source

YW, kg 0.40 480.982 1,202.455 Moser et al. (1998) UREA, sq. cm 0.29 16.501 56.900 Moser et al. (1998) UIMF, % 0.38 0.176 0.470 MacNeil and Northcutt

(2008) UFAT, cm 0.39 0.012 0.031 MacNeil and Northcutt

(2008) FI, kg 0.39 0.275 0.705 Arthur et al. (2001) HCW, kg 0.59 520.010 881.373 Moser et al. (1998) REA, sq. cm 0.39 19.008 48.738 Moser et al. (1998) FAT, cm 0.27 0.019 0.070 Moser et al. (1998) MS2_{, score 0.55 0.203 0.360 Gregory et al. (1995)}

1_{Selection criteria: YW = yearling weight, UREA = ultrasound ribeye}

area, UIMF = ultrasound intramuscular fat percentage, UFAT = ultrasound rib fat. Objective traits: FI = feed intake, HCW = hot carcass weight, REA = ribeye area, FAT = 12th–rib fat, MS = marbling score. h2 _{= heritability,}

σα2= genetic variance, σp2 = phenotypic variance.

2_{Marbling score units where 4.0 = Sl}0 _{and 5.0 = Sm}0_.

Table 5. Genetic correlations (above diagonal) between selection criteria and objective traits, and phenotypic correlations (below diagonal) between selection criteria

Trait1 _{YW UREA UIMF UFAT FI HCW REA FAT MS}

YW 0.447 _0.317 _0.033 _0.510 _0.613 _0.63 _0.322 _-0.211 UREA 0.413 _-0.257 _0.043 _0.449 _0.413 _0.663 _-0.118 _-0.38 UIMF 0.037 _-0.087 _0.367 _0.539 _0.252 _0.234 _0.336 _0.474 UFAT 0.133 _0.113 _0.177 _0.299 _0.274 _-0.246 _0.693 _0.456 FI 0.669 _0.219 _0.499 _0.59 HCW 0.123 _-0.13 _0.252 REA -0.053 _-0.212 FAT 0.352

1_{Selection criteria: YW = yearling weight, UREA = ultrasound ribeye}

area, UIMF = ultrasound intramuscular fat percentage, UFAT = ultrasound rib fat; Objective traits: FI = feed intake, HCW = hot carcass weight, REA = ribeye area, FAT = 12th–rib fat, MS = marbling score.

2_{Koots et al. (1994).} 3_{Moser et al. (1998).} 4_{Reverter et al. (2000).} 5_{Devitt and Wilton (2001).} 6_{Kemp et al. (2002).} 7_{Stelzleni et al. (2002).} 8_{Bergen et al. (2005).} 9_{Nkrumah et al. (2007).} 10_{Arthur et al. (2001).}

known without error. However, EBV would never be

known with complete certainty given the heterogene-ity of the residual variance. Thus, the index accuracy estimated herein would be the ‘best case scenario’

pre-suming that the accuracy of each EBV included in the

index for each animal was unity. Estimating Index Sensitivity

Economic selection index coefficients are seldom

known without error because of uncertainties in (co) variances and in economic values. One way to deter-mine the sensitivity of indices to the (co)variances and

economic values assumed is to calculate the efficiency of the index. The efficiency (E_u) is given as:

12 11 12 1 = u = t × t t t H u u H u u t R b' G v E R b' G b b' G v

where R_Hu is the response expected from the ‘used’ val-ues, R_Ht is the response expected from the ‘true’ values, b_u are index coefficients derived from ‘used’ values and

b_t are ‘true’ index coefficients. The ‘used’ index coeffi -cients are based on current belief, while the ‘true’ index

coefficients are assumed to be optimum. In reality, there

are potential uncertainties associated with the assumed phenotypic and genetic parameter estimates and eco-nomic values which is why it is important to calculate

the efficiency and determine the impact of inadvertently using incorrect index coefficients.

Sensitivity to absolute changes in genetic correla-tions between objective traits and selection criteria of ± 0.2 and ± 0.4 were calculated. These changes in ge-netic correlations are equivalent to those investigated by Simm et al. (1986). It is important to note that in some cases these changes resulted in a change of sign. In instances where these changes would have resulted in a correlation greater than unity, the genetic correla-tion was assumed to be 1. Sensitivity to a 50% increase or decrease in the magnitude of the economic value of each trait in the breeding objective was also investigat-ed. This also follows the methods of Simm et al. (1986),

who calculated the efficiency of 2 selection indices fol -lowing an increase or decrease of 50% in the economic value of each trait in the aggregate breeding value.

Two alternative sets of index coefficients were de -rived to test the implication of assuming half of the animals were fed through a calf-fed system while the other half were fed through a yearling system. One

set of index coefficients were calculated assuming all

calves were fed through a calf-fed system. The other

set of index coefficients were calculated assuming all

calves were fed through a yearling system. The corre-lation between these 2 indices was calculated.

RESULTS AND DISCUSSION

Economic Values

Economic values, relative economic values (REV) and the proportion of emphasis placed on each objec-tive trait are presented in Table 6. As expected, the

REV estimated for HCW, MS, and REA were positive.

Twelfth rib fat was characterized by a negative econom-ic value because increased FAT also increased numeri-cal YG, and consequently reduced the carcass value when YG exceeded 4.0. Since FI is an expense related trait, in was no surprise that the economic value for this trait was strongly negative. Hot carcass weight received 59.5% of the emphasis, implying that selection based on the index will result in the most gain in HCW. Feed intake received the next greatest emphasis at 19.3%.

Amer et al. (2001) defined 5 breeding objectives for

beef cattle in Ireland and used these to derive selection

sub-indexes for which separate sets of REV were report -ed. One of the sub-indexes proposed was a production sub-index, aimed to improve carcass value in a termi-nal objective. Breeding objective traits in the production sub-index included weaning weight (WW), winter and

summer FI (expressed in effective energy units), HCW,

carcass fat score (expressed on a 15-point scale), and carcass conformation score. To enable comparisons of results from the current study to results reported by Amer

et al. (2001), the REV were converted to US dollars us

-ing the June 2016 exchange rate. The REV of HCW was

10.38. Moreover, the relative emphasis placed on HCW was 64%, which closely aligns with the relative emphasis

placed on HCW in the present study. Amer et al. (2001)

reported REV (relative emphasis) for summer FI, winter

FI and carcass fat score of -0.20 (1%), -0.62 (4%) and -1.58 (10%), respectively. These values are consistent with results of the current study.

Barron Lopez (2013) estimated the REV of eleven

breeding objective traits aimed at improving the

ef-Table 6. Economic values, relative economic values

(REV) and relative emphasis placed on individual

objective traits

Trait1 Economic value _{($/trait unit)} Genetic SD 2

(σ_α ) (per REV σ_α) emphasis (%)Relative FI, kg -57.05 0.52 -29.66 19.3 HCW, kg 4.00 22.80 91.29 59.5 REA, sq. cm 1.92 4.36 8.38 5.5 FAT, cm -50.51 0.14 -7.07 4.6 MS, units3 _{37.80 0.45 17.01 11.1}

1_{FI = feed intake, HCW = hot carcass weight, REA = ribeye area, FAT =}

12th–rib fat, MS = marbling score.

2_{From additive genetic variances in Table 4.} 3_{Marbling score units where 4.0 = Sl}0 _{and 5.0 = Sm}0_.

Ochsner et al. 1068

ficiency of general purpose beef cattle. The objective

traits included were milk production, average daily gain, mature weight, dressing percentage, FAT,

kidney-pel-vic-heart fat, REA, MS, calving difficulty, heifer preg

-nancy, and gestation length. The REV of the carcass

traits FAT, REA, and MS were -6.90, 9.31, and 11.023, respectively. These values are very similar to those re-ported herein for the same carcass traits. Relative em-phasis placed on FAT, REA, and MS by Barron Lopez (2013) was 12%, 17%, and 20%, respectively. Again these results are comparable to results in the present study, although these three carcass traits received a pro-portionately lower percentage of the relative emphasis due to the fact that this was general purpose breeding

objective and included more traits. In comparison, REV

for carcass weight, carcass conformation score, carcass

fat score, gestation length, and calving difficulty report -ed by Amer et al. (1998) for terminal sires were 15.0, 7.3, 4.4, 3.2, and 7.8, respectively.

Buchanan et al. (2016) conducted a study evaluat-ing the economic impact of bovine respiratory disease (BRD) in a typical feedlot finishing operation. Traits

included in the breeding objective were BRD inci-dence, HCW, YG, camera marbling score, dry matter intake, days to harvest, and WW. For HCW, YG, cam-era marbling score, and dry matter intake Buchanan et

al. (2016) reported REV (relative emphasis) of 191.98

(31%), -13.59 (5%), 24.50 (9%) and -60.71 (10%), respectively. Hot carcass weight received the second highest emphasis, following BRD incidence rate. The

negative REV for YG reported by Buchanan et al. (2016) supports both the negative REV for FAT and the positive REV for REA reported herein. The REV

reported for camera marbling score is similar in

direc-tion to the REV for MS reported in the current study. The negative REV for dry matter intake is comparable

to the REV for FI derived herein.

Index Coefficients

Index coefficients for phenotypic measures of YW,

UREA, UFAT, and UIMF were 0.74, 0.08, -31.04, and

13.32, respectively. Terminal index coefficients to be applied to EBV for YW, UREA, UFAT, and UIMF

were 1.72, 0.81, -36.60, and 12.38, respectively. The correlated responses in goal traits were 0.42 kg., 21.17 kg., 4.28 cm2_{, 0.05 cm, and -0.15 units for FI, HCW,} REA, FAT, and MS, respectively.

Enns and Nicoll (2008) developed a selection in-dex for New Zealand beef cattle with an economic breeding objective aimed at increasing net income per cow lifetime. The selection criteria included WW, YW, number of calves weaned, and average lifetime body weight of calf weaned. Index weighting factors for

each trait changed depending on the number of calves

weaned by the dam. The index coefficient of YW for

a yearling out of a cow that had weaned one calf was

0.63, which is similar in sign to the index coefficient

derived for YW in the current study.

Barron Lopez (2013) estimated index coefficients for a variety of indices designed to improve the effi -ciency of general purpose beef production. In total 13 selection criteria traits were considered including av-erage daily gain, mature weight, FAT, REA, MS,

calv-ing difficulty, heifer pregnancy, birth weight, WW, YW,

HCW, yearling height, and maternal weaning weight.

The estimates of index coefficients reported for YW

ranged from 0.03 to 0.64. For the index that Barron Lopez (2013) recommended to improve the proposed

breeding objective, index coefficients for FAT, REA,

and MS were -53.0, 1.92, and 25.3, respectively. These

carcass trait index coefficients are in agreement with the index coefficients presented herein.

Index Accuracy

The accuracy of the terminal index to be used for

EBV lies between 0.338 and 0.503. The lower bound

of the accuracy estimate assumes that phenotypic mea-sures are the selection criteria. The upper bound of the

accuracy estimate assumes that EBV known without er -ror are the selection criteria. We would expect the true accuracy of the index to lie somewhere between the 2 accuracies presented herein that were produced by as-suming the index was comprised of either phenotypic

measures or by EBV that are known without error.

Index Sensitivity

The sensitivity to changes in genetic correlations

is reported as the efficiency of the index after adding

0.2 or 0.4, or after subtracting 0.2 or 0.4, from the genetic correlations between the objective traits and selection criteria, one at a time. A change of ± 0.2 in

the genetic correlations resulted in efficiencies rang -ing from 0.97 to 1.00, with the exception of

correla-tions involving HCW. Selection efficiencies resulting

from the ± 0.2 adjustment of correlations between HCW and other traits ranged from 0.85 to 0.97. The increased sensitivity of HCW to changes in its corre-lation with other traits was due to the fact that it had

the largest REV of all traits considered.

A change of ± 0.4 in the values of the correlations

re-sulted in selection efficiencies ranging from 0.94 to 1.00, with the same exception as before. Efficiencies resulting

from the adjustment ± 0.4 in genetic correlations between HCW and other traits ranged from 0.23 to 0.93. The

genetic correlation between YW and HCW, and indicates that this index is sensitive to uncertainties in genetic cor-relations between these 2 traits. To further this sensitiv-ity, 0.3 was subtracted from the ‘true’ genetic correlation

between YW and HCW; the resulting efficiency was

0.57. The genetic relationship between HCW and YW is known to be moderate to strong and positive (Koots et al., 1994). Decreasing this genetic correlation by more than 0.2 assumes a genetic relationship that is not biologically reasonable. Consequently, it can be concluded that the index is insensitive to realistic changes in the assumed genetic correlation between these 2 traits.

The sensitivity to changes in economic values is

reported as the efficiency of the index after a 50% in -crease or de-crease in the economic value of each

objec-tive trait, one at a time. Efficiency values ranged from

0.84 to 1.00. The index was the most sensitive to a 50% decrease in the economic value of HCW. The same ra-tionale applies here as for the sensitivity of HCW to changes of genetic correlation. Aside from the sensi-tivity of HCW to the decrease in economic value, all

other efficiencies calculated for the terminal index were

above 0.97. This result indicates that the index is rela-tively insensitive to wide changes in economic values.

The correlation between 2 alternative sets of index

coefficients (1 assuming all calves were fed through a

calf-fed system and the other assuming all calves were fed through a yearling system) was 0.99. Based on this

result, it can be concluded that the index coefficients

are relatively insensitive to which system was used.

Therefore, the index coefficients presented can be applied

regardless of the choice of a calf- or yearling-fed system. Conclusions

In the terminal objective considered for this study, decreasing FAT and FI while increasing HCW, REA,

and MS would increase profitability. Hot carcass weight and FI are the top 2 drivers of profit, implying that improving feed efficiency is crucial to increasing the profitability of an operation with a terminal objec -tive, more so than simply improving carcass quality

alone. Given the suite of EBV currently available to

Beefmaster breeders, this would correspond to a

selec-tion index based on EBV with positive coefficients for

yearling weight, and ultrasonically measured REA and intramuscular fat percentage. Ultrasonically measured

FAT EBV would have a negative index coefficient.

Multitrait selection is critical given that more than

one trait impacts overall profitability of a beef cattle op

-eration. The most efficient way to conduct multitrait se -lection is by using an economic se-lection index. Although the index from the current study was proven to be robust to changes in the assumed genetic correlations and

eco-nomic values, care should be taken relative to the appli-cation of the index in production systems that might vary in terms of production goals. In example, the terminal index assumed herein does not contemplate the retention of females for the purposes of breeding. Using the index from the current study in a commercial production sce-nario where breeding females are retained would not be

advised given the potential differences in goal traits. For

a terminal objective in Beefmaster herds, selection based on the economic selection index presented herein would

improve the profitability of commercial beef enterprises.

LITERATURE CITED

Amer, P. R., R. Crump, and G. Simm. 1998. A terminal sire se-lection index for UK beef cattle. J. Anim. Sci. 67:445–454. doi:10.1017/S1357729800032859

Amer, P. R., G. Simm, M. G. Keane, M. G. Diskin, and B. W. Wickman. 2001. Breeding objectives for beef cattle in Ireland. Livest. Prod. Sci. 67:223–239. doi:10.1016/S0301-6226(00)00201-3 Arthur, P. F., J. A. Archer, D. J. Johnston, R. M. Herd, E. C.

Richardson, and P. F. Parnell. 2001. Genetic and phenotypic variance and covariance components for feed intake, feed

efficiency, and other postweaning traits in Angus cattle. J.

Anim. Sci. 79:2805-2811. doi:10.2527/2001.79112805x Barron Lopez, J. A. 2013. Selection Objective for Improving

Efficiency of Beef Cattle. PhD Diss. Univ. of Nebraska, Lincoln. Bergen, R., S. P. Miller, and J. W. Wilton. 2005. Genetic

corre-lations among indicator traits for carcass composition

mea-sured in yearling beef bulls and finished feedlot steers. Can. J. Anim. Sci. 85:463–473. doi:10.4141/A05-013

Bourdon, R. M. 1998. Shortcomings of current genetic evaluation sys-tems. J. Anim. Sci. 76:2308–2323. doi:10.2527/1998.7692308x Buchanan, J. W., M. D. MacNeil, R. C. Raymond, A. R. McClain, and

A. L. Van Eenennaam. 2016. Variance component estimates for

Charolais-sired fed cattle and the relative economic impact of bo-vine respiratory disease. J. Anim. Sci. doi:10.2527/jas.2016-1001 Devitt, C. J., and J. W. Wilton. 2001. Genetic correlation estimates

between ultrasound measurements of yearling bulls and

car-cass measurements on finished steers. J. Anim. Sci. 79:2790– 2797. doi:10.2527/2001.79112790x

Enns, R. M., and G. B. Nicoll. 2008. Genetic change results from selection on an economic breeding objective in beef cattle. J. Anim. Sci. 86:3348–3357. doi:10.2527/jas.2006-566

Falconer, D. S., and T. F. C. Mackay. 1996. Introduction to Quantitative

Genetics, Ed 4. Longmans Green, Harlow, Essex, UK. Forabosco, F., A. Groen, R. Bozzi, J. Van Arendonk, F. Filippini, P.

Boettcher, and P. Bijma. 2004. Phenotypic relationships between longevity, type traits, and production in Chianina beef cattle. J. Anim. Sci. 82:1572–1580. doi:10.2527/2004.8261572x

Gregory, K. E., L. V.Cundiff, and R. M. Koch. 1995. Genetic and phenotypic (co)variances for growth and carcass traits of pure-bred and composite populations of beef cattle. J. Anim. Sci. 73:1920–1926. doi:10.2527/1995.7371920x

Hazel, L. N. 1943. The genetic basis for constructing selection indexes. Genetics 28:476–490.

Hazel, L. N., and J. L. Lush. 1942. The efficiency of three methods

of selection. J. Hered. 33:393–399.

Hill, W. G. 1974. Prediction and evaluation of response to selec-tion with overlapping generaselec-tions. Anim. Prod. 18:117–139. doi:10.1017/S0003356100017372

Ochsner et al. 1070

Kemp, D. J., W. O. Herring, and C. J. Kaiser. 2002. Genetic and en-vironmental parameters for steer ultrasound and carcass traits. J. Anim. Sci. 80: 6: 1489-1496. doi:10.2527/2002.8061489x Koots, K. R., J. P. Gibson, and J. W. Wilton. 1994. Analyses of

published genetic parameter estimates for beef production traits. 2. Phenotypic and genetic correlations. Anim. Breed. Abstr.62:825–853.

Livestock Marketing Information Center (LMIC). 2015. http:// www.lmic.info/. (Accessed 10 August 2015.)

MacNeil, M. D., and S. L. Northcutt. 2008. National cattle evalu-ation system for combined analysis of carcass characteristics and indicator traits recorded by using ultrasound in Angus cattle. J. Anim. Sci. 86:2518–2524. doi:10.2527/jas.2008-0901 Moser, D. W., J. K. Bertrand, I. Misztal, L. A. Kriese, and L. L.

Benyshek. 1998. Genetic Parameter Estimates for Carcass and Yearling Ultrasound Measurements in Brangus Cattle. J. Anim.

Sci. 76:2542–2548. doi:10.2527/1998.76102542x

Murphey, C. E., D. K. Hallett, W. E. Tyler, and J. C. Pierce, Jr. 1960. Estimating yields of retail cuts from beef carcasses. J. Anim. Sci. 19:1240 (Abstr.).

Nkrumah, J. D., J. A. Basarab, Z. Wang, C. Li, M. A. Price, E. K. Okine, D. H. Crews, Jr., and S. S. Moore. 2007. Genetic and phenotypic relationships of feed intake and measures of

ef-ficiency with growth and carcass merit of beef cattle. J. Anim. Sci. 85:2711–2720. doi:10.2527/jas.2006-767

Ponzoni, R. W., and S. Newman. 1989. Developing breeding ob-jectives for Australian beef cattle production. Anim. Prod. 49:35–47. doi:10.1017/S0003356100004232

Reverter, A., D. J. Johnston, H. U. Graser, M. L. Wolcott, and W. H. Upton. 2000. Genetic analysis of live-animal ultrasound and abat-toir carcass traits in Australian Angus and Hereford cattle. J. Anim. Sci. 78:1786–1795. doi:10.2527/2000.7871786x

Rolfe, K. M., W. M. Snelling, M. K. Nielsen, H. C. Freetly, C. L. Ferrell, and T. G. Jenkins. 2011. Genetic and phenotypic parameter estimates for feed intake and other traits in grow-ing beef cattle, and opportunities for selection. J. Anim. Sci. 89:3452–3459. doi:10.2527/jas.2011-3961

Schneeberger, M., S. A. Barwick, G. H. Crow, and K. Hammond. 1992. Economic indices using breeding values

pre-dicted by BLUP. J. Anim. Breed. Genet. 109:180–187. doi:10.1111/j.1439-0388.1992.tb00395.x

Simm, G., C. Smith, and J. H. D. Prescott. 1986. Selection indices to

improve the efficiency of lean meat production in cattle. Anim. Prod. 42:183–193. doi:10.1017/S000335610001789X

Stelzleni, A. M., T. L. Perkins, A. H. Brown, Jr., F. W. Pohlman, Z. B. Johnson, and B. A. Sandelin. 2002. Genetic parameter estimates of yearling live animal ultrasonic measurements in Brangus cattle. J. Anim. Sci. 80:3150–3153. doi:10.2527/2002.80123150x USDA-AMS. 2015. http://www.ams.usda.gov/. (Accessed 6 July,

2016.)

USDA-NASS. 2015. https://quickstats.nass.usda.gov/. (Accessed 24 June, 2015.)

Van Vleck, L. D. 1993. Selection index and introduction to mixed mod -el methods for genetic improvement of animals: The green book. CRC Press, Inc. Boca Raton, FL.

Wheeler, T. L., L. V. Cundiff, L. D. Van Vleck, G. D. Snowder, R.

M. Thallman, S. D. Shackelford, and M. Koohmaraie. 2006.

Preliminary results from cycle VIII of the cattle germplasm

evaluation program at the Roman L. Hruska U.S. Meat Animal Research Center. Germplasm Evaluation Program Progress Report No. 23. Clay Center, NE.